It's an Android app, I'm doing some async network ...

# coroutines

[kirillrakhman]

It's an Android app, I'm doing some async network requests and then I'm manipulation the UI. The network requests can throw `IOException`s or `HttpException`s so I need to catch them, otherwise the app will crash.

[elizarov]

Some you can catch just `IOException`s, can't you?

[kirillrakhman]

and HttpException, at least (multi catch would be useful)

[kirillrakhman]

but I'm also catching other exceptions and submit them to my crash reporting instead of letting the app crash

[elizarov]

And how do you handle your IOExceptions? Like some specific code everywhere, so you’ve encapsulated the way you handle IOExceptions into some common code?

[kirillrakhman]

I have some generic retry logic for the httprequests (I'm using Rx for this) but if it fails after n times, I just swallow it and set the UI to some fail state

[kirillrakhman]

@elizarov so if (e is CancellationException) throw e is ok in my catch block?

[elizarov]

I see thanks. Yes, it is Ok. You can also have the same line in your crash reporter.

[kirillrakhman]

ok, thanks for help. do you think there is a way, to streamline this?

[elizarov]

I would extract into some some common code. Actually, I’d install a crash reporter as CoroutineExceptionHandler https://kotlin.github.io/kotlinx.coroutines/kotlinx-coroutines-core/kotlinx.coroutines.experimental/-coroutine-exception-handler/index.html in my context and make it ignore CancellationException, but report everything else

[kirillrakhman]

but where do I put the UI code then?

[elizarov]

Basically, the idea is that instead of doing launch(UI) { … } to start a coroutine you do luanch(myOwnContext) { … }, where myOwnContext = UI + myExceptionHandler

[cbruegg]

@kirillrakhman How do you deal with this?

[cbruegg]

https://github.com/Kotlin/kotlinx.coroutines/issues/47

[kirillrakhman]

@cbruegg currently I don't

[kirillrakhman]

i.e. I have automatic cancelation based on Navi, but I don't check the lifecycle after every suspension point

[kirillrakhman]

@elizarov I understand, but I still need to know when the coroutine fails, so I need to extend my exception logger context and provide that?

[elizarov]

It is hard to get a more specific advice without looking at your actual code

[cbruegg]

Are you sure though your jobs are actually canceled even when they're already in the Handler queue? (If you don't know what I mean and have the time, feel free check out the example project I've linked to)